Logistic regression or something else?

[D3mand]

I have a dependent binary outcome and 6 independent variables. These are measured in the same group of people at two moments in time in a descriptive longitudinal study. I am assessing whether these independent variables increase the likelihood of achieving the binary outcome. I know that I could do a logistic regression analysis of the results gathered at either one of those moments in time but I was wondering if there was an analysis I could perform that would incorporate both measurements of the same participants at those two points in time.

 

[obh]

Hi D3,

Using any regression to choose the model is a bit problematic, as you have multiple comparisons.
has any of the IV values changed for a specific participant between the two moments in time?

 

[D3mand]

Hi obh,

Thank you for your reply. No, none of the independent values have changed, but their effect on the dependent variable may have changed if that makes sense! By running a regression model would it not allow me to determine the overall fit of the model and the relative contribution of each of the predictors to the total variance explained? Since I have two data points for each participant in a longitudinal study, I was hoping that each of these could work as their own control and make my results more accurate. I was thinking a mixed-effects model might work best, but I am open to suggestions.

 

[SoManyQuestions]

I have a dependent binary outcome and 6 independent variables. These are measured in the same group of people at two moments in time in a descriptive longitudinal study. I am assessing whether these independent variables increase the likelihood of achieving the binary outcome. I know that I could do a logistic regression analysis of the results gathered at either one of those moments in time but I was wondering if there was an analysis I could perform that would incorporate both measurements of the same participants at those two points in time.

By "incorporate both measurements of the same participants at those two points in time", do you mean include all the observations in a single model? If so, I would include an indicator equal to one if the observation is after the treatment, and then interact that with the IV's.

lmao I just realized how old this post is, nvm